Quote:
Originally Posted by deolio
To that point - just today I was in a meeting with a client whom had been reading a book they got for Christmas and she said a couple pages in something felt "off". After googling the book, it turned out to be one of those made-to-order Amazon books written by an unheard of company selling hundreds of books on countless unrelated topics all with the same generic design treatment. In the reviews of the book, a number of people had come to the same conclusion as her and felt that the book (and all of the others by that company) was purely copy and pasted from ChatGPT with zero editing. A bit concerning if you ask me...
|
The thing that kind of cemented my opinion of these generative AIs is what you get if you ask it a question like "how many of the letter n appear in the word mayonnaise" and it gives an extremely confident-sounding but utterly nonsensical answer. The ultimate danger of AI to society is not that the robots will take over, but that people think the robots are
smart when it fact the robots are just
really good at bullshitting.
Now when you give it a task like "describe this product using these keywords and specs" "generate an image based on all the other similar images of this style" "write some example code that does this" it's pretty good. Though the first time I threw a code problem at chatgpt I stumped it - maybe it's my own fault for not being able to prompt it correctly, but after several tries it just kept giving me the textbook solution to what I was asking and not the slightly odd, 'anti-pattern' thing I was asking it to do.